W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > May to August 1996

Re: no-transform

From: Ted Hardie <hardie@merlot.arc.nasa.gov>
Date: Thu, 6 Jun 1996 09:01:43 -0700 (PDT)
Message-Id: <199606061601.JAA08715@merlot.arc.nasa.gov>
To: jg@w3.org
Cc: fielding@liege.ICS.UCI.EDU, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/826
Jim writes: 
> Here is an attempt to rewrite this:
> 14.9.5 No-Transform Directive
>  Intermediate caches (proxies) have found it useful to convert the
>  media type of an entity-body under some circumstances. For example, a
>  proxy might use convert image formats to other, more compact formats
>  or sizes before transmission on a slow link, or to save cache space.
>  HTTP has to date been silent on these transformations, but serious
>  operational problems are now known to occur frequently for certain
>  kinds of applications, which have failed as a result.
>  Some entity-bodies must not undergo any transformation before being provided
>  to applications.  Examples include: medical imagery, scientific data
>  and data using end-to-end authentication, where it is important that
>  the original entity-body be exactly reproduced bit for bit in the
>  application.  Therefore, an intermediate cache MUST NOT change the
>  media type of an entity-body if the response includes the no-transform
>  directive.

I like the new direction of the language, as I think it makes clear
the conditions under which a no-transform might be applied and why
it must be obeyed. 

The first paragraph is a little hard to parse, though.  How about
the following as a re-write:

  Intermediate caches (proxies) have found it useful to convert the
  media type of certain entity bodies.  A proxy might, for example,
  convert between image formats in order to save cache space or
  to reduce the amount of traffic on a slow link.  HTTP has to date
  been silent on these transformations.  Serious operational problems
  do occur, however, when these transformations are applied to entity
  bodies which are part of certain kinds of applications, and these
  applications can fail as a result.

I think this language still fits with the second paragraph and is
little easier to follow.

				Ted Hardie
Received on Thursday, 6 June 1996 09:06:00 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:40:17 UTC